Our community wishes to understand the access patterns of our wiki. There
seems to be quite a few special pages and extensions that tell us how
many edits occur (both globally and individually). However, after looking
around a little bit, I could find no software that would analyze how many
pages are accessed on a daily basis by readers (not editors).
Software such as Webalizer count web page access using the squid/apache
log files. But those logs include entries that cannot be considered
actual content accesses. For example, our apache logs are densely
littered with lines such as:
... "GET /skins/common/commonPrint.css?270 HTTP/1.0" ...
... "GET /skins/common/shared.css?270 HTTP/1.0" ...
... "GET /skins/Pinkwich5/main.css?270 HTTP/1.0" ...
... "GET /skins/common/mwsuggest.js?270 HTTP/1.0" ...
... "GET /skins/common/ajax.js?270 HTTP/1.0" ...
...
...
... "GET /wiki?title=MediaWiki:Common.css ...
... "GET /wiki?title=MediaWiki:Pinkwich5.css ...
... "GET /wiki?title=MediaWiki:Print.css
These GET requests do not represent content page accesses.
I am looking for software that counts page accesses and has the
capability to eliminate page edits and perhaps even limit counts to non-
logged in users. I also need software that will automatically generate
reports that are (automatically) embedded in the wiki (using a Special
page or other approach).
--
-- Dan Nessett